Process Capability Index Calculator (Cp, Cpk, Cpm, Pp, Ppk)
Enter your specification limits, process mean, and standard deviation to calculate the full suite of process capability indices: Cp, Cpk, Cpm, Pp, and Ppk. You also get the sigma level, expected defect rate in parts per million (PPM), and a plain-English verdict on whether your process is capable. Results update instantly as you type.
What is the process capability index?
Process capability indices are dimensionless numbers that compare what a manufacturing or business process actually produces to what it is supposed to produce, as defined by the specification limits (USL and LSL). A value of 1.0 means the process just fits inside the spec at 3-sigma - about 2,700 defective parts per million. A value of 1.33 (4-sigma) means 64 PPM, and a value of 2.0 (6-sigma) means fewer than 0.002 PPM. The indices are used everywhere from automotive stamping to pharmaceutical tablet weight to software response-time SLAs.
Cp vs. Cpk: potential vs. actual capability
Cp measures potential capability - how wide the spec is relative to the process spread, assuming the process is perfectly centered. It answers the question: "Could this process be capable if we centered it?" Cpk is the real-world index: it takes the smaller of the upper and lower capability ratios, so it shrinks whenever the mean drifts away from the center of the specification band. A process with Cp = 2.0 but Cpk = 0.8 has plenty of room in the spec but the mean is sitting far to one side, causing many failures. The gap between Cp and Cpk tells you whether your improvement effort should focus on reducing variation (raise Cp) or re-centering the process (raise Cpk toward Cp).
Cpm, Pp, and Ppk - when to use them
Cpm (the Taguchi index) modifies Cp by replacing the standard deviation with a combined measure of variation and distance from a nominated target value T. When your target is at the center of the spec, Cpm equals Cp. When the process drifts away from T, Cpm penalises this more severely than Cpk, making it popular in industries where hitting a specific nominal is critical, not just staying within limits. Pp and Ppk are the long-term equivalents of Cp and Cpk: they use the overall standard deviation across all subgroups rather than the within-subgroup estimate. A large gap between Cpk and Ppk (say, more than 0.2) indicates that the process shifts or drifts over time, which a control chart would help detect.
How to interpret and act on your results
If Cpk is above 1.33 but Cp is also close to Cpk, the process is centered and capable - maintain it with a control chart. If Cp is much higher than Cpk, shift the process mean toward the specification midpoint: this is usually cheaper and faster than reducing variability. If both Cp and Cpk are below 1.0, the process variation is too large for the given tolerance: you must reduce common-cause variation through designed experiments (DOE), better material control, or equipment calibration. If Cpk is between 1.0 and 1.33, the process is marginal: check whether the distribution is truly normal (skewed data can give misleading index values) and consider tightening monitoring frequency.
Cpk Interpretation Guide
| Cpk | Sigma Level | PPM (approx.) | Verdict | Typical Requirement |
|---|---|---|---|---|
| < 1.00 | < 3 | > 2,700 | Not capable | Immediate action required |
| 1.00 - 1.33 | 3 - 4 | 64 - 2,700 | Marginal | Monitor closely; improve |
| 1.33 - 1.67 | 4 - 5 | 0.6 - 64 | Capable | Automotive / general industry minimum |
| 1.67 - 2.00 | 5 - 6 | 0.002 - 0.6 | Excellent | Aerospace / medical devices |
| >= 2.00 | >= 6 | < 0.002 | World-class | Six Sigma target |
Industry-standard thresholds used in Six Sigma, automotive (AIAG), and aerospace quality standards.
Frequently asked questions
What is a good Cpk value?
The widely accepted minimum is Cpk = 1.33 (4-sigma, about 64 PPM). Most automotive standards (AIAG PPAP) require at least 1.67 for critical characteristics. Aerospace and medical-device standards often require 2.0 or higher. A value below 1.0 means the process is not capable and is producing defects at a rate above 2,700 PPM.
What is the difference between Cpk and Ppk?
Cpk uses the within-subgroup (short-term) standard deviation, which reflects only common-cause variation within a single run. Ppk uses the overall (long-term) standard deviation across all subgroups and time periods, so it also captures process drift and shifts. Cpk tells you what the process could do if it were perfectly stable; Ppk tells you what it actually did. A healthy process has Ppk close to Cpk.
Why is Cpk lower than Cp?
Cp assumes the process is perfectly centered midway between the LSL and USL. Cpk uses the actual process mean and takes the smaller of the two one-sided capability ratios. Whenever the mean is off-center, Cpk is lower than Cp. The difference quantifies how much capability is being lost to off-centering rather than to inherent variation.
What does a negative Cpk mean?
A negative Cpk means the process mean has moved outside one of the specification limits. For example, if the mean is above the USL, the upper ratio (USL - mean)/(3 sigma) is negative. The process is producing a majority of output outside the specification - immediate intervention is required.
Do capability indices require a normal distribution?
Yes - the standard Cp and Cpk formulas assume the process output follows a normal distribution. If the data is skewed or multi-modal, the indices can give misleading results. For non-normal processes, either apply a transformation (e.g., Box-Cox) to achieve normality, or use a non-normal capability analysis based on percentile methods.
How many data points do I need for a reliable Cpk estimate?
A minimum of 30 data points is commonly cited for a rough estimate, but 100 or more gives a Cpk confidence interval narrow enough to be actionable. With fewer points, use the lower confidence bound of Cpk (often called Cpk lower) for decision-making rather than the point estimate.